Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [RE] Updated paper



Fred,

You raise some good points; much appreciated.

I believe you point can be illustrated in this example:
  1) High quality VOIP negotiates for 1ms interval,
     because it needs the low latency. But, in actuality,
     it only sends every 8ms.
  2) Video at 8ms interval is impacted, since the VOIP
     "appears" to take 8x the bandwidth.

I think small modifications to the originally conceived
subscription protocols can handle this case. In this
example, the consumed BW could be "reclaimed" by the
lower rate transmissions, so that VOIP only "appears"
to take a 1x BW, for traffic at the 8ms rate.

Since this affects only subscription (not the frame
forwarding/shaping/pacing), the added complexity is
expected to be minimal.

Thanks again for your most helpful comments.

DVJ



>> -----Original Message-----
>> From: Tuck, Fred [mailto:Fred.Tuck@echostar.com]
>> Sent: Wednesday, September 07, 2005 3:19 PM
>> To: 'David V James'; STDS-802-3-RE@listserv.ieee.org
>> Subject: RE: [RE] Updated paper
>>
>>
>> David:
>>
>> In section 10.1 of your rate based scheduling proposal you equate shorter
>> latency with higher bandwidth utilization (or at least reservation).  It
>> seems to me that in many application areas used in the home that
>> the latency
>> time needed may actually be inversely proportional to the
>> bandwidth.  High
>> bandwidth applications like HD video need a bounded latency to ensure
>> adequate buffer sizes and proper delay before decoding starts but the
>> absolute value of the latency is not as important.  Some low bandwidth
>> applications, especially interactive ones like corroborative music, need
>> very low latencies to properly sync all of the performers.  It
>> would seem a
>> waste to reserve a lot of bandwidth for the low bandwidth
>> applications that
>> need low latency that would then preclude higher bandwidth
>> applications from
>> running due to maximizing the reservation limits.
>>
>> Fred Tuck
>> EchoStar
>>
>> -----Original Message-----
>> From: David V James [mailto:dvj@ALUM.MIT.EDU]
>> Sent: Wednesday, August 31, 2005 1:19 AM
>> To: STDS-802-3-RE@listserv.ieee.org
>> Subject: [RE] Updated paper
>>
>> All,
>>
>> I have posted an update for tomorrow's meeting.
>> This can be found at:
>>   http://dvjames.com/esync/dvjRate2005Aug30.pdf
>>
>> I assume Michael will move this to the group's
>> IEEE web pages.
>>
>> DVJ